Research On Tree Branch Modeling Method Of L-system Based On Image Sequence

3D tree model is an important part of virtual geographic environment.The current three-dimensional tree modeling shows the trend from landscape representation to describing growth.A tree morphogenetic model suitable for dynamic development plays a fundamental role in the establishment of three-dimensional tree growth model.L-system is a common model of tree morphogenesis,which can simulate the geometric and topological structure of real trees on the one hand,and can be coupled with the physiological and ecological model on the other hand,so it has been widely used.However,due to the difficulty of data acquisition and the complexity of the rule generation process of L system,it is difficult to extract the L system rules of real tree branches quickly and truly.To solve this problem,this paper designs a tree branch segment model based on the real tree branch point cloud and supports the generation of grammar rules of parameterized random L system.On this basis,the method of model construction and the method of automatic generation of rules of parameterized random L system are proposed.The research contents and achievements of this paper mainly include:(1)Design of tree branch data model for L systemThrough the interpretation of L system principle and model related dendrology principle,the basic structural unit of "branch section" is extracted,and the model parameters of branch section object are summarized.The function of the model is realized in four stages: "real tree-tree branch point cloud-tree branch skeleton line-tree branch segment".The conceptual model is established for the data objects involved in the integrated process,and the logical expression model and physical storage model are given.Compared with the traditional tree structure model,the model can not only store the morphological data of real tree branches,but also build a bridge between real tree branches and parameterized random L-system rules.(2)Construction of tree branch model based on image sequenceTake enough tree image sequences around the target trees,and generate tree point clouds from the image sequences.On this basis,generate skeleton lines of tree branches,extract tree branches from the skeleton as the basic unit of parametric random L-system rule extraction,measure and calculate the key parameters of the model,and fill them into the model database.The experiment shows that the topological structure of the tree segment model constructed by this method is the same as that of the real tree,and the accuracy of the important parameter attribute data is higher than that of the measured data,which is a branch segment model construction method worthy of reference.(3)The automatic generation method of L system rulesThis paper analyzes the result of tree segment model,divides it into the combination of tree segment monomer,generates axiom,production formula and instantiation rule of parameter for target tree branches,edits segment of special segment and outlier,and realizes the automatic generation of L system rule.Experiments show that the automatic rule generation method is convenient and effective,which can restore the geometry and topology of tree branches better,and the branch editing method can be used as a supplement,adding artificial intervention to achieve the purpose of modifying the details of tree branches.Through the discussion,L-system method can be used to build tree branches with strong ductility,which can realize the accurate positioning of texture adding and organs,coupling with plant physiological and ecological models,and differential batch modeling.In this paper,a method based on image sequence is proposed to model the real tree branches.In this method,the geometric topological features of tree branches are extracted from the image sequence,and the feature attributes are filled in the model which links the real tree geometric topological parameters with the parameterized random L system.On this basis,the grammar rules of the parameterized random L system are automatically generated,and the branches are edited and visualized.The results show that this method can generate the L-system rules of tree branches and trunks realistically and efficiently,and can be used in large-scale scene differential tree modeling,which lays a foundation for the rapid and accurate establishment of threedimensional tree growth model with realistic shape and in line with the physiological process of trees,and makes basic work for the completion of the construction of virtual geographic environment.